Final Project

#load required packages
library(tidyverse)
library(leaflet)
library(sp)
library(sf)
library(rgdal)

#Data Manipulation

#load data
df <- read_csv('jan_dec.csv')

Parsed with column specification:
cols(
  .default = col_double(),
  tpep_pickup_datetime = col_datetime(format = ""),
  tpep_dropoff_datetime = col_datetime(format = ""),
  store_and_fwd_flag = col_character()
)
See spec(...) for full column specifications.

#replace Inf values with 0
df <-  do.call(data.frame,lapply(df, function(x) replace(x, is.infinite(x),0)))

Error in as.POSIXct.numeric(value) : 'origin' must be supplied

#replace Inf values with 0
df <- df %>%
  mutate(avg_mph = ifelse(avg_mph == Inf, 0, avg_mph))

#Data by Pickup Zone

df_pu <- df %>%
  group_by(PULocationID) %>%
  summarize(passenger_count = mean(passenger_count),
            fare_amount = mean(fare_amount),
            trip_distance = mean(trip_distance),
            extra = mean(extra),
            mta_tax = mean(mta_tax),
            tip_amount = mean(tip_amount),
            tolls_amount = mean(tolls_amount),
            total_amount = mean(total_amount),
            fare_by_dist = mean(fare_by_dist),
            duration = mean(duration),
            avg_mph = mean(avg_mph),
            adj_total = mean(adj_total))

#Within Zone Data

df_within <- df %>%
  filter(PULocationID == DOLocationID) %>%
    group_by(PULocationID) %>%
  summarize(passenger_count = mean(passenger_count),
            fare_amount = mean(fare_amount),
            trip_distance = mean(trip_distance),
            extra = mean(extra),
            mta_tax = mean(mta_tax),
            tip_amount = mean(tip_amount),
            tolls_amount = mean(tolls_amount),
            total_amount = mean(total_amount),
            fare_by_dist = mean(fare_by_dist),
            duration = mean(duration),
            avg_mph = mean(avg_mph),
            adj_total = mean(adj_total))

#Pickup Volume Data

df_volume <- df %>%
  group_by(PULocationID) %>%
  tally()

taxi_zones <- readOGR("taxi_zones/taxi_zones.shp")

OGR data source with driver: ESRI Shapefile 
Source: "/Users/athvedt/Documents/GitHub/Data Visualization/Group_W/taxi_zones/taxi_zones.shp", layer: "taxi_zones"
with 263 features
It has 6 fields

#transform polygon
proj <- spTransform(taxi_zones, '+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs')

Map 1 - Average Tip by Pickup Zone

content <- paste("Neighborhood:", taxi_zones$zone, "<br/>",
                 "Average Tip:", "$", round(df_pu$tip_amount, digits = 2), "<br/>")

bins <- c(0, 1, 2, 3, 4, 5, 15)
pal <- colorBin("Greens", domain = df_pu$tip_amount, bins = bins)

leaflet(df_pu) %>%
  addTiles() %>%
  setView(lng = -73.98928, lat = 40.75042, zoom = 10.2) %>%
  addProviderTiles("CartoDB.Positron") %>%
  addPolygons(data = proj,
              popup = content,
              weight = 1,
              fillColor = ~pal(df_pu$tip_amount),
              fillOpacity = 1,
              highlightOptions = highlightOptions(
                color='#000000',
                weight = 3,
                bringToFront = TRUE,
                sendToBack = TRUE),
              label = taxi_zones$zone) %>%
  addLegend("topright",
            pal = pal,
            values = df_pu$tip_amount,
            title = "Average Credit Card Tip (USD)",
            opacity = 1,
            labFormat = labelFormat(prefix = "$"))

Map 2 - Average Speed

#Issues - No data for certain zones, shows “NA” on map

content <- paste("Neighborhood:", taxi_zones$zone, "<br/>",
                 "Average Speed (MPH):", round(df_within$avg_mph, digits = 5), "<br/>")


leaflet(df_within) %>%
  addTiles() %>%
  setView(lng = -73.98928, lat = 40.75042, zoom = 10.2) %>%
  addProviderTiles("CartoDB.Positron") %>%
  addPolygons(data = proj,
              popup = content,
              weight = 1,
              fillColor = ~colorQuantile("RdYlGn", df_within$avg_mph)(df_within$avg_mph),
              fillOpacity = 1,
              highlightOptions = highlightOptions(
                color='#000000',
                weight = 3,
                bringToFront = TRUE,
                sendToBack = TRUE),
              label = taxi_zones$zone) %>%
  addLegend("topright",
            pal = colorQuantile("RdYlGn", df_within$avg_mph, n = 5),
            values = df_within$avg_mph,
            title = "Average Speed (Percentile)",
            opacity = 1,)

Map 3 - Pickup Volume

content <- paste("Neighborhood:", taxi_zones$zone, "<br/>",
                 "Number of Pickups", df_volume$n, "<br/>")


leaflet(df_volume) %>%
  addTiles() %>%
  setView(lng = -73.98928, lat = 40.75042, zoom = 10.2) %>%
  addProviderTiles("CartoDB.Positron") %>%
  addPolygons(data = proj,
              popup = content,
              weight = 1,
              fillColor = ~colorQuantile("YlOrRd", df_volume$n)(df_volume$n),
              fillOpacity = 1,
              highlightOptions = highlightOptions(
                color='#000000',
                weight = 3,
                bringToFront = TRUE,
                sendToBack = TRUE),
              label = taxi_zones$zone) %>%
  addLegend("topright",
            pal = colorQuantile("YlOrRd", df_volume$n, n = 5),
            values = df_volume$n,
            title = "Pickup Volume Percentile",
            opacity = 1,)